Retrieval Augmented Generation Based LLM Evaluation For Protocol State Machine Inference With Chain-of-Thought Reasoning

This paper presents a novel approach to evaluate the efficiency of a RAG-based agentic Large Language Model (LLM) architecture for network packet seed generation and enrichment. Enhanced by chain-of-thought (COT) prompting techniques, the proposed approach focuses on the improvement of the seeds' structural quality in order to guide protocol fuzzing frameworks through a wide exploration of the protocol state space. Our method leverages RAG and text embeddings to dynamically reference to the Request For Comments (RFC) documents knowledge base for answering queries regarding the protocol's Finite State Machine (FSM), then iteratively reasons through the retrieved knowledge, for output refinement and proper seed placement. We then evaluate the response structure quality of the agent's output, based on metrics as BLEU, ROUGE, and Word Error Rate (WER) by comparing the generated packets against the ground-truth packets. Our experiments demonstrate significant improvements of up to 18.19%, 14.81%, and 23.45% in BLEU, ROUGE, and WER, respectively, over baseline models. These results confirm the potential of such approach, improving LLM-based protocol fuzzing frameworks for the identification of hidden vulnerabilities.